2. Insurance Coverage, Provider Contact, and Take-Up of the HPV Vaccine
2.7. Appendix
Table A1: Summary statistics of additional variables
(1) (2) (3)
Overall Expansion
States
Non-Expansion States
Male 0.511 0.510 0.512
(0.500) (0.500) (0.500)
Age Indicators
14 0.198 0198 0.199
(0.399) (0.698) (0.399)
15 0.207 0.207 0.208
(0.406) (0.405) (0.406)
16 0.206 0.204 0.207
(0.404) (0.403) (0.405)
17 0.188 0.192 0.186
(0.391) (0.394) (0.389)
Grade-Level Indicators
6-8th 0.274 0.259 0.283
(0.446) (0.438) (0.451)
9-12th 0.713 0.728 0.703
(0.452) (0.445) (0.457)
High School Graduate 0.009 0.009 0.009
(0.096) (0.096) (0.097)
Race/Ethnicity Indicators
White 0.551 0.524 0.568
(0.497) (0.499) (0.495)
Black 0.141 0.105 0.163
(0.348) (0.307) (0.370)
Hispanic 0.216 0.259 0.189
(0.411) (0.438) (0.391)
Mother’s Age Indicators
≤ 34 0.093 0.079 0.102
(0.291) (0.269) (0.303)
35-44 0.444 0.415 0.462
(0.497) (0.493) (0.499)
Mother’s Education Indicators
< High School 0.127 0.137 0.121
(0.333) (0.343) (0.326)
High School Graduate 0.235 0.220 0.245
(0.424) (0.414) (0.430)
123
Some College 0.258 0.242 0.268
(0.486) (0.428) (0.443)
Household Income Indicators
< $20K 0.190 0.181 0.196
(0.392) (0.385) (0.397)
$20-30K 0.107 0.102 0.111
(0.310) (0.303) (0.314)
$30-40K 0.085 0.080 0.088
(0.279) (0.272) (0.283)
$40-50K 0.074 0.067 0.078
(0.261) (0.250) (0.269)
Time-Varying State Controls
Unemployment Rate 6.839 6.171 7.264
(2.363) (2.185) (2.373)
Tdap Mandate 0.899 0.970 0.854
(0.301) (0.172) (0.353)
Meningococcal Mandate 0.427 0.511 0.374
(0.495) (0.500) (0.484)
Source: National Immunization Survey—Teen 2010-2018
124
Table A2: Summary statistics for HPV vaccine initiation that do not utilize the sample weights
(1) (2) (3) (4) (5) (6) (7)
Full Sample
Expansion States Non-Expansion States
All Years 2010 2018 All Years 2010 2018
Mean 0.484 0.506 0.248 0.713 0.451 0.237 0.636
Standard Deviation (0.500) (0.500) (0.432) (0.453) (0.498) (0.425) (0.481)
Observations 172,891 104,254 10,859 10,375 68,637 7,100 7,331
Source: National Immunization Survey—Teen 2010-2018
Note: HPV initiation is an indicator for whether provider-verified immunization records indicate that the child had received at least one dose of the HPV vaccine.
125
Table A3: Across specifications, the event study specification does not find any relationship between Medicaid expansion and HPV vaccination in the pre-period and a positive relationship in the post-period
(1) (2) (3) (4) (5)
Pre-Expansion
-4 -0.000 -0.018 0.002 0.003 -0.050
(0.015) (0.014) (0.017) (0.014) (0.032)
-3 -0.014 -0.025* -0.010 -0.007 -0.041*
(0.015) (0.013) (0.015) (0.014) (0.024)
-2 -0.011 -0.023* -0.009 -0.010 -0.029*
(0.013) (0.012) (0.014) (0.012) (0.015) Pre=0?
F-Stat 0.530 1.670 0.460 0.390 1.290
Prob>F 0.665 0.187 0.711 0.758 0.288
Post-Expansion
0 0.019 0.005 0.020 0.016 0.028*
(0.013) (0.015) (0.014) (0.012) (0.015) 1 0.034** 0.025* 0.034** 0.030** 0.054**
(0.015) (0.014) (0.016) (0.013) (0.022) 2 0.052*** 0.042** 0.051*** 0.047*** 0.083**
(0.018) (0.021) (0.018) (0.015) (0.032) Post=0?
F-Stat 3.800 2.690 3.560 3.910 2.350
Prob>F 0.016 0.057 0.021 0.014 0.084
Pre=Post?
F-Stat 2.290 2.250 2.870 3.050 1.930
Prob>F 0.022 0.065 0.023 0.018 0.106
State and Year FE? Y Y Y Y Y
Demographic Controls? N N Y Y Y
State-Level Covariates? N N N Y Y
State-Specific LTT? N N N N Y
Including Early Expanders? Y N Y Y Y
Observations 172,891 157,987 172,891 172,891 172,891
Source: National Immunization Survey 2010-2018
Note: The dependent variable is an indicator for receiving at least one dose of the HPV vaccine. The independent variables are indicator variables for being j periods away from Medicaid expansion. Column (1) includes controls for time-invariant state fixed effects and location-invariant year fixed effects. Column (2) uses this same
specification but excludes states which expanded Medicaid prior to 2014 as part of the ACA. Column (3) includes demographic controls, column (4) state-level policies, and column (5) state-specific linear time trends. The exact controls are detailed in Table 2.
*** p<0.01, ** p<0.05, * p<0.10
126
Table A4: Medicaid expansion was unrelated to whether an observation had adequate provider information
(1) (2)
Observations with HPV Vaccine Information
All Observations
Medicaid Expansion -0.006 -0.002
(0.006) (0.007)
Mean 0.979 0.497
Observations 176,536 309,830
Note: The dependent variable is an indicator for whether the teen has adequate provider-verified vaccination data. The independent variable of interest is an indicator for whether the state expanded Medicaid as part of the Affordable Care Act. All columns include the full set of controls from Table 2 column (4). Columns (1) examines only observations with data on HPV vaccination, while column (2) analyzes all observation. Robust standard errors, shown in parentheses, are clustered at the state level.
*** p<0.01, ** p<0.05, * p<0.10
127
Table A5: Most of the difference-in-differences estimate is identified from comparing treated states to never treated states
(1) (2)
Weight Avg DD Estimate
Earlier Treated vs. Later Control 0.067 0.032 Later Treated vs. Earlier Control 0.101 0.021
Treated vs. Never Treated 0.718 0.043
Treated vs. Already Treated 0.113 -0.038
Source: National Immunization Survey 2010-2018
Note: The dependent variable is the share of teens vaccinated against HPV, while the independent variable of interest is an indicator for whether the state had expanded Medicaid as part of the ACA. The regression includes state and year fixed effects. The weight assigned to each comparison group and the average difference-in-differences coefficient is obtained from first collapsing the data to the state-year level and then using the bacondecomp command with ddetail.
128
Table A6: The relationship between Medicaid expansion and HPV vaccination is driven by comparing changes in expansion states to changes in non-expansion states, as opposed to comparing changes in early- and late-expanders
(1) (2) (3) (4)
Excluding States which Expanded but not in 2014
Excluding States which Expanded but not in 2014
Only Leveraging the 2014 ACA
Expansions
Only Leveraging the 2014 ACA
Expansions
Medicaid Expansion 0.038*** 0.030 0.030*** 0.031*
(0.013) (0.019) (0.011) (0.016)
State & Year FE? Y Y Y Y
Full Set of Covariates? N Y N Y
Mean 0.474 0.474 0.487 0.487
Observations 136,532 136,532 172,891 172,891
Source: National Immunization Survey 2010-2018
Note: The dependent variable is an indicator for whether the child’s immunization provider reports that the child had received at least one dose of the HPV vaccine. The independent variable of interest is an indicator for whether the state expanded Medicaid as part of the Affordable Care Act. Odd numbered columns includes only time- invariant state fixed effects and location-invariant year fixed effects. Even numbered columns include the full set of controls from Table 2 column (4). Columns (1) and (2) exclude states which expanded Medicaid as part of the ACA in any year except 2014. Columns (3) and (4) uses all observations but redefines treatment as an indicator which takes on the value of 1 if the state expanded Medicaid in 2014 and 0 otherwise. Robust standard errors, shown in parentheses, are clustered at the state level.
*** p<0.01, ** p<0.05, * p<0.10
129
Table A7: The relationship between Medicaid expansion and HPV vaccine initiation was not statistically different for teen boys and teen girls
(1) (2) (3) (4)
Boys Girls Full
Sample
Full Sample 2012-2018
Medicaid Expansion 0.040* 0.026 0.027 0.003
(0.022) (0.018) (0.017) (0.020)
Medicaid Expansion * Boy 0.012 0.024
(0.026) (0.029)
Mean 0.599 0.379 0.487 0.547
Observations 90,431 82,460 172,891 132,893
Source: National Immunization Survey—Teen 2010-2018
Note: The dependent variable is an indicator for whether the child’s immunization provider reports that the child had received at least one dose of the HPV vaccine. The independent variable of interest is an indicator for whether the state expanded Medicaid as part of the Affordable Care Act. Column (1) restricts the sample to teen boys and column (2) restricts the sample to teen girls. Columns (3) and (4) use a triple-difference specification whereby every covariate is interacted with an indicator for being male. Column (4) further restricts the sample to the years 2012-2018 after ACIP recommended the HPV vaccine for teen boys. Each column includes the full set of controls from Table 2 column (4), and the estimates utilize the sample weights. Robust standard errors, shown in parentheses, are clustered at the state level.
*** p<0.01, ** p<0.05, * p<0.10
130
Table A8: Medicaid expansion was associated with an increase in the probability of parent-reported HPV vaccine take-up for poorer teens, teens whose mothers lacked college degrees, and non-white teens
(1) (2) (3) (4) (5) (6) (7)
Excluding Responses from Shotcard
≤ 200%
FPL
> 200%
FPL
Mother lacked BA
Mother had BA
Non-
White White
Medicaid Expansion 0.017* 0.029** 0.015 0.028*** 0.011 0.032** 0.006
(0.010) (0.012) (0.013) (0.009) (0.015) (0.013) (0.011)
Mean 0.420 0.444 0.413 0.417 0.438 0.459 0.400
Observations 282,379 92,567 194,102 154,164 132,595 102,014 184,745
Source: National Immunization Survey—Teen 2010-2018
Note: The dependent variable is an indicator for whether the child’s parent reports that the teen had received at least one dose of the HPV vaccine. The independent variable of interest is an indicator for whether the state expanded Medicaid as part of the Affordable Care Act.
Column (1) does not include observations from parents who used a shotcard to answer the question in years 2010-2013. Instead, it exclusively uses recall responses. Columns (2)-(7) utilize the full sample of parent-reported vaccination. Columns (2) and (3) stratify the sample by poverty status. Similarly, columns (4) and (5) stratify the sample by mother’s education, and columns (6) and (7) by race/ethnicity. Each column includes the full set of controls from Table 2 column (4), and the estimates utilize the sample weights. Robust standard errors, shown in parentheses, are clustered at the state level.
*** p<0.01, ** p<0.05, * p<0.10
131
Table A9: Medicaid expansion was associated with greater public health insurance coverage for poorer individuals and teens whose mothers lacked college degrees
(1) (2) (3) (4) (5) (6) (7)
Full Sample
≤ 200%
FPL
> 200%
FPL
Mother lacked BA
Mother had BA
Non-
White White Panel II: Public Health Insurance
Medicaid Expansion 0.014* 0.026* 0.005 0.017 0.007 0.018 0.006
(0.008) (0.013) (0.006) (0.011) (0.005) (0.013) (0.006)
Mean 0.374 0.696 0.110 0.513 0.134 0.531 0.246
Panel II: Non-Public Health Insurance
Medicaid Expansion -0.001 -0.008 0.004 -0.001 0.000 0.002 0.001
(0.006) (0.012) (0.006) (0.008) (0.006) (0.011) (0.006)
Mean 0.563 0.200 0.860 0.402 0.842 0.377 0.715
Observations 198,169 70,006 128,163 110,044 88,125 71,808 126,361 Source: National Immunization Survey—Teen 2010-2018
Note: The dependent variable in Panel I is an indicator for whether the child was covered by public health insurance. The dependent variable in Panel II is an indicator for whether the child was covered by any health insurance that was not public insurance. The independent variable of interest is an indicator for whether the state expanded Medicaid as part of the Affordable Care Act. Column (1) examines the full sample, while columns (2) and (3) stratify the sample by poverty status.
Similarly, columns (4) and (5) stratify the sample by mother’s education, and columns (6) and (7) by race/ethnicity. Each column includes the full set of controls from Table 2 column (4), and the estimates utilize the sample weights. Robust standard errors, shown in parentheses, are clustered at the state level.
*** p<0.01, ** p<0.05, * p<0.10
132
Table A10: The relationship between Medicaid expansion and health insurance coverage is robust to only examining the 2010-2015 period prior to a survey modification limiting which observations include information about health insurance coverage
(1) (2) (3) (4) (5) (6) (7)
Full Sample
≤ 200%
FPL
> 200%
FPL
Mother lacked BA
Mother had BA
Non-
White White Medicaid Expansion 0.018*** 0.026** 0.011*** 0.021*** 0.010* 0.022** 0.012**
(0.006) (0.010) (0.004) (0.007) (0.005) (0.009) (0.005)
Mean 0.929 0.883 0.967 0.905 0.973 0.894 0.957
Observations 141,258 50,192 91,066 80,524 60,734 50,355 90,903
Source: National Immunization Survey—Teen 2010-2015
Note: The dependent variable is an indicator for whether the child was covered by any health insurance. The independent variable of interest is an indicator for whether the state expanded Medicaid as part of the Affordable Care Act. Column (1) examines the full sample, while columns (2) and (3) stratify the sample by poverty status. Similarly, columns (4) and (5) stratify the sample by mother’s education, and columns (6) and (7) by race/ethnicity. Each column includes the full set of controls from Table 2 column (4), and the estimates utilize the sample weights. Robust standard errors, shown in parentheses, are clustered at the state level.
*** p<0.01, ** p<0.05, * p<0.10
133
Table A11: Medicaid expansion was associated with a reduction in the probability that a parent gives “lack of knowledge” as a reason for the child not receiving the HPV vaccine within the subsequent 12 months
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
Not Needed
No Rec.
Lack Knowledge
Safety
Concerns Cost Not
Needed
No Rec.
Lack Knowledge
Safety
Concerns Cost Pre-Expansion
-4 -0.010 0.011 -0.006 -0.015 0.017* 0.000 0.083** 0.018 -0.027 0.009
(0.008) (0.014) (0.008) (0.009) (0.009) (0.031) (0.039) (0.026) (0.024) (0.012)
-3 -0.018* 0.015 -0.010 -0.002 0.006 -0.012 0.066** 0.006 -0.012 -0.001
(0.009) (0.014) (0.010) (0.012) (0.008) (0.021) (0.032) (0.016) (0.022) (0.007)
-2 -0.014* -0.011 0.007 0.004 0.012* -0.016 0.023 0.016 0.002 0.005
(0.007) (0.014) (0.011) (0.010) (0.007) (0.013) (0.016) (0.015) (0.015) (0.004) Pre=0?
F-Stat 2.010 2.130 0.770 1.860 3.250 2.080 1.650 0.560 2.780 2.550
Prob>F 0.124 0.108 0.514 0.148 0.029 0.114 0.190 0.643 0.051 0.066
Post-Expansion
0 0.011 -0.002 -0.010 -0.008 0.009** -0.003 -0.009 -0.013 -0.001 0.007
(0.010) (0.009) (0.008) (0.007) (0.004) (0.015) (0.011) (0.011) (0.007) (0.006)
1 -0.013 -0.001 -0.030*** 0.011* 0.014** -0.031 -0.023 -0.039*** 0.024** 0.012
(0.014) (0.009) (0.009) (0.006) (0.007) (0.031) (0.014) (0.014) (0.009) (0.009)
2 -0.013 -0.006 -0.011 0.009 0.012* -0.036 -0.038* -0.024 0.020 0.012
(0.013) (0.010) (0.009) (0.010) (0.006) (0.034) (0.022) (0.017) (0.013) (0.012) Post=0?
F-Stat 0.980 0.180 4.650 3.560 2.000 0.770 1.400 4.440 4.890 0.710
Prob>F 0.410 0.907 0.006 0.021 0.125 0.519 0.253 0.008 0.005 0.550
Pre=Post?
F-Stat 2.280 1.800 8.400 2.950 1.170 0.940 1.090 3.830 4.390 1.340
Prob>F 0.060 0.130 0.000 0.021 0.338 0.463 0.380 0.005 0.002 0.263
Observations 126,395 126,395 126,395 126,395 126,395 126,395 126,395 126,395 126,395 126,395
Source: National Immunization Survey 2010-2018
Note: The dependent variable is an indicator for the reason given for not vaccinating the child, including that the vaccine is not needed; the child has not been recommended the vaccine; a lack of knowledge; safety concerns; and the cost of the vaccine. The independent variables are indicator variables for being j periods away from Medicaid expansion.
Columns (1)-(5) include controls for time-invariant state fixed effects and location-invariant year fixed effects. Columns (6)-(1) use the full set of controls from Table 2 column (4).
The estimates utilize the sample weights. Standard errors, shown in parentheses, are clustered at the state level.
*** p<0.01, ** p<0.05, * p<0.10
134
Figure A1: After accounting for CDC funds to promote HPV vaccination, HPV vaccination rates were trending similarly in states which did and did not expand Medicaid as part of the ACA
(A) (B)
(C) (D)
Source: National Immunization Survey- Teen 2010-2018
Note: The figures plot the unweighted share of teen girls and teen boys who had initiated the HPV vaccine by whether the teens resided in states which expanded Medicaid as part of the ACA. Panels (A) and (C) presents the estimates for boys and girls, respectively. Panels (B) and (D) present these same shares but exclude states which received CDC funding to improve HPV vaccination.
135 CHAPTER 3
E-Verify Mandates and Unauthorized Immigrants’
Health Insurance Coverage
3.1. INTRODUCTION
Immigrants disproportionately rely on the labor market for health insurance due to restrictions on public insurance for new authorized arrivals and unauthorized immigrants (Borjas 2003). Nearly a quarter of lawful permanent residents and over 40 percent of unauthorized immigrants lack health insurance (KFF 2017), and over the last two decades state and local governments have experimented with policies intended to disrupt unauthorized immigrants’ access to the formal labor market. One such policy is the requirement that at least some employers electronically verify (E- Verify) that their new hires are eligible to work in the United States.
E-Verify mandates previously enjoyed bipartisan support (Politico 2013), and several high- profile GOP leaders have expressed optimism about achieving a comprehensive immigration reform package with the Biden administration (The Hill 2020). Because E-Verify mandates remain popular throughout the Republican party (White House 2017; White House 2018; Romney 2019) a nationwide mandate is almost certain to be a part of these discussions. Indeed, a nationwide E- Verify mandate was included in proposed legislation from Senators Mitt Romney (R-UT) and Tom Cotton (R-AR) to increase the federal minimum wage to $10 an hour (Vox 2021). Yet President Biden has simultaneously called for expanding unauthorized immigrants’ access to health insurance, going so far as to propose allowing unauthorized immigrants to enroll in public health insurance plans (Washington Post 2019). In absence of such a proposal, barring unauthorized immigrants from the formal labor market through a nationwide E-Verify mandate could eliminate
136 their only option for health insurance coverage.
In this paper, I show that state-level E-Verify mandates reduced the probability that likely unauthorized immigrants had health insurance coverage, a relationship driven by reductions in the likelihood of having employer-sponsored insurance. Event study estimates demonstrate the reduction only occurred after the mandate was implemented, and in a series of falsification tests I show that that naturalized citizens, Hispanic natives, and white non-Hispanic natives did not experience a similar change. Interestingly, the effect for likely unauthorized immigrants was limited to the period immediately after implementation. In all subsequent periods, the relationship between E-Verify mandates and health insurance coverage was near zero. I show that this pattern can be explained by selective outmigration of otherwise unemployed and subsequently uninsured likely unauthorized immigrants.
This paper contributes to the literature on immigrants’ access to insurance by demonstrating a plausibly causal link between E-Verify mandates and health insurance coverage (Borjas 2003; Buchmueller et al. 2008; Bronchetti 2014; Dillender 2017). Additionally, it adds to a growing body of work on the effect of E-Verify mandates on likely unauthorized immigrants which has thus far focused primarily on employment outcomes and migration decisions (Amuedo- Dorantes and Bansak 2014; Bohn et al. 2014; Orrenius and Zavodny 2015; Orrenius and Zavodny 2016). It also contributes to a broader literature on the spillover effects of immigration enforcement (Bitler and Hoynes 2011; Watson 2014; Amuedo-Dorantes, Arenas-Arroyo, and Sevilla 2018; East 2019; Churchill, Amuedo-Dorantes, and Song 2020) by showing suggestive evidence that children with likely unauthorized parents and native adults in mixed-status households lose access to health insurance.
The rest of this paper proceeds as follows: Section 2 discusses existing work on E-Verify,
137
as well as the literature on immigrants and health insurance, discussing how unauthorized immigrants can obtain insurance coverage. The data, methods, and summary statistics are discussed in Section 3. Section 4 starts by showing that E-Verify mandates reduce likely unauthorized immigrants’ employment prospects, an effect which is driven by reduction in wage- employment and employment at larger firms. I then show that this effect is limited to one period post-implementation, after which point unauthorized immigrants opt to leave the state. I then present the main insurance results. Finally, Section 5 discusses broad conclusions and opportunities for future work.
3.2. EXISTING LITERATURE
Since 2007, nine states have implemented laws requiring all employers to utilize E-Verify, and an additional fourteen states require public employees or contractors to be screened through E-Verify.
Proponents argue that these mandates can reduce the flow of unauthorized immigrants (or induce return migration), while also benefitting citizen workers. For instance, Congressman Lamar Smith (R-Texas) stated, “E-Verify is the most effective deterrent to illegal immigration because it shuts off the jobs magnet and saves jobs for hardworking Americans” (CNN 2018).
3.2.1. E-Verify and Employment
The Immigration Reform and Control Act of 1986 barred firms from knowingly hiring or employing unauthorized immigrants. However, uneven enforcement (Reyes et al. 2002) did little to stem the flow of unauthorized labor into the United States (Amuedo-Dorantes and Bansak 2014). A decade later, the Illegal Immigration Reform and Immigrant Responsibility Act of 1996 established the Basic Pilot program. Now known as E-Verify, this program compares information from a new hire’s Form I-9 against databases maintained by the Social Security Administration
138
and Department of Homeland Security, helping employers assure they hire authorized workers (Stumpf 2012). E-Verify was made available to select states beginning in 1997, with all states having access by 2003 (Orrenius and Zavodny 2015).
There is mixed evidence on the relationship between E-Verify mandates and unauthorized immigrants’ labor market outcomes. Focusing on Arizona’s E-Verify mandate, Bohn and Lofstrom (2012) found reductions in wage-and-salary employment for non-citizen Hispanics. Examining a broader set of universal and public E-Verify mandates with the 2004-2011 Current Population Survey (CPS) data, Amuedo-Dorantes and Bansak (2014) also found employment reductions for likely unauthorized immigrants and improved job prospects for those competing with unauthorized labor. However, when using the 2002-2012 CPS data, Orrenius and Zavodny (2015) failed to detect a negative employment effect; indeed, their point estimate was positive and statistically insignificant.
E-Verify mandates may also affect state composition by (i) inducing unauthorized immigrants to leave the state and/or (ii) discouraging future unauthorized immigrants from settling.
Bohn, Lofstrom, and Raphael (2014) found Arizona’s E-Verify mandate reduced the fraction of the state’s population comprised of Hispanic non-citizens. Looking at a broader group of states, Orrenius and Zavodny (2016) found that universal E-Verify mandates reduced the number of likely unauthorized immigrants in a state. While they found evidence that unauthorized immigrants settled in other states in response to E-Verify laws, they also posited that some unauthorized individuals may have opted to return to their native countries. On the other hand, using administrative data from the Department of Homeland Security on the usage of E-Verify systems, Ayromloo, Feigenberg, and Lubotsky (2020) did not find evidence that these mandates induced work-ineligible individuals to relocate.
139 3.2.2. Immigrants and Health Insurance
The 1996 Personal Responsibility and Work Opportunity Reconciliation Act (PRWORA) barred lawful permanent residents (LPRs) from most means-tested programs during their first five years in the US. Borjas (2003) found that the PRWORA eligibility changes did reduce Medicaid participation, though affected immigrants compensated by increasing their labor supply to gain employer-sponsored insurance, which indicates the existence of crowd-out. However, several papers suggest that there is less crowd-out for immigrant children (Currie 2000; Kaushal and Kaestner 2005, 2007; Lurie 2008).
Under PRWORA, states had the option to offer LPRs public insurance, though they were barred from using federal money for this purpose until 2002. After this point, limited funds were available for prenatal care through the SCHIP “unborn child” option, and these funds were expanded in 2009 through the SCHIP reauthorization bill (Bitler and Hoynes 2011). Bronchetti (2014) examined these state actions to restore access to public health insurance and found that expanded eligibility increased take-up of public insurance among immigrant children.
In addition to reductions expected mechanically from changes in eligibility, there is a growing awareness that hostile policy environments may exacerbate reductions in program take- up (Fix and Passel 1999; Borjas 2001; Kandula et al. 2004). For example, the PRWORA-induced reductions in Medicaid participation could not be entirely explained by eligibility changes, leading Borjas (2003) to attribute the disproportionate response to chilling effects. Sommers (2010) found that the Deficit Reduction Act (DRA) of 2005, which imposed citizenship documentation requirements on Medicaid applicants, reduced the share of adult immigrants enrolled in Medicaid, though the overall adult insurance rate was not affected.
140
There is also evidence that some unauthorized immigrants forgo health care visits due to fears of interacting with law enforcement officers (Núñez and Heyman 2007; Heyman et al. 2009).
Watson (2014) found that increased federal immigration enforcement lowered Medicaid participation among children with immigrant mothers while also decreasing (increasing) the probability that these children were reported to be in Very Good Health (Poor Health). Similarly, Alsan and Yang (2018) found that county participation in the Secure Communities program reduced the probability that a Hispanic citizen utilized means-tested benefit programs, such as SNAP and SSI.
Given these restrictions on public insurance, immigrants must largely rely on private health insurance. However, immigrants are less likely to have private insurance relative to their native counterparts, in part because they are less likely to be employed by firms offering health insurance coverage. Indeed, Buchmueller et al. (2007) found that the citizen/noncitizen coverage gap could largely be explained by noncitizens working at firms which did not offer employer-sponsored health insurance. Among those working at firms offering health insurance, noncitizens were only slightly less likely to be eligible for coverage and, among that group, only slightly less likely to take up coverage. Building off this finding, Dillender (2017) showed that immigrants possessing stronger English ability were more likely to have employer-sponsored health insurance. These barriers are especially acute for unauthorized immigrants. Unauthorized immigrants are barred from receiving the Affordable Care Act’s private insurance subsidies. While it is possible to obtain coverage outside the Marketplace or through an employer without providing a Social Security Number, the cost is often prohibitive (KFF 2019).
3.3. DATA, MEASURES, AND METHODS
I first obtained preliminary information on state E-Verify mandates from the National Council of